Zurück

What are tokens?

A quick introduction to how tokens work when using Astell

What are tokens?

Tokens measure metered usage across Astell: Processed Data ingestion (documents, images, audio, video, meetings, web pages, and people/company entities), chat (every model), Astell Actions, and the twice-daily Audits. Native Data syncing and search never consume tokens.

Which features don't consume tokens?

These features are unlimited and always available, subject only to the Fair Use Policy:

• Native Data syncing (messages, emails, tasks, code)
• Search across your connected data
• The Loops dashboard (viewing loops, opening loop detail, citations)
• Team collaboration features

(Chat, on any model, draws from your token pool. See What costs tokens.)

Which features consume tokens, and what are the rates?

• Image ingestion: 1 token / image
• Audio ingestion: 2 tokens / minute
• Document ingestion: 5 tokens / page
• Video ingestion: 8 tokens / minute
• Meeting ingestion: 10 tokens / minute
• Web page ingestion: 5 tokens / page
• People / Company entity ingestion: 25 tokens / entity

What token allocation is included in each plan?

Astell has four plans, Sapling, Tree, Grove, and Forest (Enterprise):

• Sapling: free, 2,500 tokens/month, single seat (~500 document pages). Standard (Flash) model only, 3 months of indexed history. Best for solo users who want to find the right info fast.
• Tree: $20/seat/month billed annually ($25 monthly), 10,000 tokens/seat/month (~2,000 document pages), 18 months of indexed history. Best for teams who need shared answers and richer media.
• Grove: $80/seat/month billed annually ($100 monthly), 50,000 tokens/seat/month, 60 months of indexed history. Best for larger orgs with complex permissions and workflows, plus Ultra mode.
• Forest (Enterprise): custom pricing and a customizable indexing window; contact founders@labtwofour.com.

On Tree and above, tokens are pooled at the organization level. Customized admin control over token consumption is available on Grove and Forest.

How far back does Astell index my data?

Each plan indexes a different depth of history (the "indexing window"): Astell processes data within that window, and older content falls outside it.

• Sapling: 3 months
• Tree: 18 months
• Grove: 60 months (5 years)
• Forest (Enterprise): customizable

A longer window means more of your historical context is searchable and available to your loops.

How do pooled tokens work across an organization and multiple seats?

Starting on Tree, tokens are pooled at the organization level. There is one shared token balance for the org, and all seats draw from it. Your organization's monthly tokens = (plan tokens per seat) × (number of seats). Example: 10 users on Tree = 100,000 tokens/month. Because tokens are pooled, any user's Processed Data ingestion draws from the shared balance; Grove and Forest add customized admin controls over who can consume tokens.

What is a token limit?

A token limit is a monthly cap you set on how many tokens your organization can consume in a billing month, including any pay-as-you-go overage. It is the simplest way to control spend: set it to your included monthly pool to prevent pay-as-you-go charges, or set it higher to let ingestion continue beyond the pool while still capping the maximum bill. If you hit the limit, Processed Data ingestion may pause until tokens reset or the limit is increased.

What happens when my organization runs out of tokens?

Viewing never breaks; acting needs tokens. When you hit your monthly limit:

• Native Data syncing and Search keep working normally, they never consume tokens.
• The Loops dashboard, loop detail, citations, unified search, and calendar view stay accessible. They read from the existing graph and don't require new token spend.
• Chat (every model) pauses until your tokens reset, you increase your token limit (allowing pay-as-you-go), or you upgrade your plan.
• Processed Data ingestion pauses. Anything already ingested stays searchable; only new ingestion is affected.
• Astell Actions pause. The Agent or chat can still propose what it would do, but the action isn't pushed until you have budget.

Notifications fire at 75%, 90%, and 100% of your limit.